Personality controls in settings

vendor/agent-harness/agent-workspace/AGENTS.md

@@ -10,6 +10,7 @@ On a **new chat**, recall who you are talking to before you speak:
 1. Use `USER.md` if it has their name.
 2. If the name is missing, `read_file` `USER.md` and `MEMORY.md` (and today’s `memory/YYYY-MM-DD.md` if present). Then greet them by name.
 3. If those files still have no name, ask once and write `USER.md`.
+4. Follow `PERSONA.md` for user-authored acting notes. Settings is the source of truth for that file.
 
 Do not invent a name. Skills: only the catalog is inlined. `read_file` the matching `SKILL.md` when a task fits.
 

vendor/agent-harness/agent-workspace/PERSONA.md

@@ -0,0 +1,3 @@
+# PERSONA.md
+
+User-authored notes for how the assistant should act. Settings → Voice → How they should act is the source of truth. Leave this empty if there are no extra notes.

vendor/agent-harness/agent-workspace/SOUL.md

@@ -42,6 +42,7 @@ Each chat session you wake up fresh. These workspace files _are_ you:
 - `IDENTITY.md`: name and face
 - `AGENTS.md`: how you operate
 - `USER.md`: who you’re helping
+- `PERSONA.md`: extra acting notes from Settings
 - `MEMORY.md` and `memory/YYYY-MM-DD.md`: what you’ve learned
 
 Read them. Update them when something durable happens. That is how you persist.

vendor/agent-harness/agent/policy.js

@@ -19,6 +19,7 @@ const IDENTITY_FILES = new Set([
   'bootstrap.md',
   'heartbeat.md',
   'tools.md',
+  'persona.md',
 ]);
 
 function isIdentityPath(filePath) {

vendor/agent-harness/agent/prompts.js

@@ -26,12 +26,14 @@ const VOICE_WORKSPACE_FILES = [
   'BOOTSTRAP.md',
   'TOOLS.md',
   'HEARTBEAT.md',
+  'PERSONA.md',
 ];
 
 function includeWorkspaceFile(name, text) {
   const body = String(text || '').trim();
   if (!body) return false;
   if (name === 'BOOTSTRAP.md' && /^# completed/i.test(body)) return false;
+  if (name === 'PERSONA.md' && !body.replace(/^# PERSONA\.md\s*/i, '').trim()) return false;
   return true;
 }